SAIC Motor (China)

To address the issues of insufficient multi-scale representation of features, inadequate physical periodicity, and meteorological noise interference in photovoltaic power prediction, the MT-Transformer model is proposed. Multi-scale convolution (MSC) is introduced to capture instantaneous fluctuations and evolutionary trends, periodic enhancement coding (CPE) is used to inject daily and seasonal physical priors, and dynamic gating of meteorological variables is achieved through a feature selection layer (FSL). Experiments show that the MT-Transformer performs excellently on a dataset of 12 solar stations, with an RMSE reduction of approximately <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{1 2 {\%}}$</tex> compared to the standard Transformer.

&lt;div class="section abstract"&gt;&lt;div class="htmlview paragraph"&gt;The electro-mechanical brake (EMB) system is a novel dry-type brake-by-wire system that features superior control performance and a compact structural design, effectively meeting the development demands of intelligent and electrified vehicles. However, current research on anti-lock braking system (ABS) primarily focuses on hydraulic brake system and mostly remains at the simulation and hardware-in-the-loop testing stages. Therefore, this paper validates the feasibility of slip ratio control based on EMB actuators through both simulation and real-vehicle experiments. First, this paper establishes an equivalent second-order response model for the closed-loop EMB control system through theoretical derivation and identifies the dynamic response characteristics of the EMB actuator via sinusoidal frequency sweep testing. Next, it compares two control strategies: one that uses the reference slip ratio as the direct control target, and another that uses reference wheel speed as the direct control target to indirectly regulate slip ratio. The latter effectively avoids the nonlinearities in slip ratio control caused by variations in vehicle speed. Based on reference wheel speed control, three types of slip ratio controllers were designed and derived: proportional-integral control (PI), integral sliding mode control (ISM), and super-twisting integral sliding mode control (STISM). Finally, simulation and real-vehicle tests on high-adhesion road surfaces verified that sliding mode slip ratio control based on reference wheel speed offers robustness, avoids the risks associated with overestimated controller gains, and improves the overall stability of the control system. In particular, the STISM, as a representative of high-order sliding mode control, effectively addresses the chattering issue present in traditional first-order sliding mode methods, offering enhanced braking safety and comfort.&lt;/div&gt;&lt;/div&gt;

Active suspension systems have gained significant attention for their capability to improve vehicle dynamics and energy efficiency. However, achieving consistent control performance under diverse and uncertain road conditions remains challenging. This paper proposes a prior-guided residual reinforcement learning framework for active suspension control. The approach integrates a Linear Quadratic Regulator (LQR) as a prior controller to ensure baseline stability, while an enhanced Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm learns the residual control policy to improve adaptability and robustness. Moreover, residual connections and Long Short-Term Memory (LSTM) layers are incorporated into the TD3 structure to enhance dynamic modeling and training stability. The simulation results demonstrate that the proposed method achieves better control performance than passive suspension, a standalone LQR, and conventional TD3 algorithms.

Abstract This paper presents the development of a local communication test system for the STC32G microcontroller, focusing on the Controller Area Network (CAN) and Local Interconnect Network (LIN). The STC32G, known for its robust features and reliability, is widely used in automotive electronics. The system employs Qt for efficient management of test procedures and real-time monitoring of results. We detail the architecture of the test system, the testing methodologies employed, and the results obtained from the evaluation of the STC32G’s local communication capabilities.

Lane detection is crucial for autonomous driving. Recent advancements have expanded traditional two-dimensional (2D) lane detection to three-dimensional (3D) by predicting lane positions in 3D space. These methods rely on fully supervised learning, requiring high-quality 3D labels, which are difficult to obtain. This challenge motivates a weakly supervised approach leveraging abundant and easily scalable 2D lane annotations. Specifically, we systematically analyze lane geometric structure priors and introduce Lane Geometry Supervision (LGS), which relies solely on 2D lane labels. Extensive experiments on both synthetic and real-world datasets demonstrate that our method achieves performance comparable to fully supervised approaches using direct 3D labels. Moreover, incorporating LGS as a regularization term further enhances the performance of existing fully supervised methods. Finally, we show that LGS enables a label-efficient training methodology for 3D monocular lane detection, effectively utilizing both scarce yet complete 3D lane labels and abundant but incomplete 2D lane labels.

The Bootloader based on the UDS protocol is the most commonly used for the software update in automobiles. The Bootloader needs to be fully tested to ensure it running Steadily on vehicle. This paper takes the CAN bus as an example, analyzes the relevant protocols of Bootloader, summarizes the testing requirements of Bootloader, and then develops an automated testing system. Moreover, this system has the characteristics of high automation, easy configuration, and strong readability of the test report.

In the past, the relationship between classical mechanics and relativity focused on the quantitative aspect, and it was considered that relativity described the precise quantitative relationship of things, while classical mechanics was only an approximation in the case of low speed and weak gravitational field. This paper points out that there is a special material connection between classical mechanics and relativity, where the special substance is the physical vacuum, which is everywhere, and people are always looking at the world through the physical vacuum. In the case of low speed and weak gravitational field, the density fluctuation of physical vacuum is small, and the classical mechanics without considering the effect of physical vacuum can be established; in the case of high speed and strong gravitational field, the density fluctuation of physical vacuum is large, and the relativistic effect appears, which can be seen as a “lens effect” caused by changes in the density of the physical vacuum. The so-called relativistic factor is actually the compressible factor of ether density changing with velocity. Understanding this material connection will reveal the limitations of relativity and provide new ideas for the further development of physics.